Anthropic maps three AI futures for 2030; the most extreme could upend the economy

A new framework explores how AI could affect GDP, unemployment, and wages, causing anything from modest gains to 15% annual GDP growth and nearly 1 in 5 cognitive workers unemployed.

computerworld.com

Cisco bundles fixes for multiple vulnerabilities, some critical, into one patch

The patch addresses two critical vulnerabilities and five other high-severity bugs impact Cisco’s carrier-grade routing OS, with potential paths to RCE and unauthorized access.

csoonline.com

ChatGPT, Claude, and Grok all went down at once; enterprises need a backup plan

Thursday's outages across the three services highlight an uncomfortable reality: Enterprises are automating work faster than they're planning for AI failures.

cio.com

Dell’s $95B AI backlog shows the infrastructure crunch is far from over

The company is struggling to keep pace with the infrastructure demands of the agentic AI boom, with shortages now stretching across servers, storage, and other critical infrastructure components.

networkworld.com

Google’s Gemini 3.8 Flash is built for agents, while its Cyber twin hunts vulnerabilities

Google's 3.8 Flash Cyber model enhances vulnerability detection and patching, outperforming larger models at lower costs, crucial for cybersecurity teams.

venturebeat.com

Anthropic makes changes to stop AI agents running amok again

Following three recent security incidents involving Claude, Anthropic is moving to more isolated environments, continuous monitoring, explicit instructions, and proactive human intervention.

csoonline.com

Broadcom says that enterprise AI agents need two things: Data they can trust and boundaries they can’t cross

The Tanzu Platform now combines AI-ready data infrastructure with sandboxed agents, isolated credentials, curated data products, and observability tools.

infoworld.com

ServiceNow patches three maximum severity flaws that could put enterprise data at risk

The vulnerabilities can be exploited remotely without user interaction; security teams should also look beyond ServiceNow to the credentials, APIs and workflows the platform can access.

csoonline.com

Enterprises can measure AI usage, but the hard part is proving that it actually delivered value

Tempo’s new Workforce Intelligence offering ties AI activity and cost to individual Jira work items to measure its contribution alongside that of human labor.

infoworld.com

AI is turning the semiconductor market into a $1.6 trillion industry

Gartner has raised its 2026 semiconductor forecast, with revenue now expected to jump 92%. Memory alone will account for more than half of industry growth.

networkworld.com

One in five enterprises can't stop a runaway AI agent's spending in real time

New VentureBeat Pulse Research finds enterprises run three AI orchestration platforms at once, yet 21% still can't stop runaway agent spending in real time.

venturebeat.com

When AI explains its decision, humans may stop thinking independently

Researchers found that adding LLM-generated rationales to recommendations can suppress productive human disagreement and cause evaluators to reject high-potential ideas.

cio.com

AI’s attribution problem gets worse as models scale

MIT researchers found that as diffusion models scale, individual training examples have less measurable influence on outputs, and this “attribution decay” could complicate AI copyright, auditing, and ...

computerworld.com

AI inference is getting cheaper, but your agents are getting more expensive

As agents reason, replan, call other agents, and work continuously in the background, Gartner predicts inference costs per workflow will rise more than fivefold through 2028.

cio.com

DeepSeek's top-ranked V4 Flash stumbles on real agent tasks as its prices surge

DeepSeek's V4 Flash tops AI leaderboards but completed just 53.8% of real-world agent tasks in a new test — as the company hikes API prices up to 1,100%.

venturebeat.com

DeepSeek raises some V4 prices by more than 10x as AI demand strains capacity

DeepSeek is raising its prices, to the dismay of developers, but off-peak pricing, cache discounts, and multi-model routing mean the impact is more nuanced.

computerworld.com

AI agents are turning data silos into an existential infrastructure problem

Separate studies conducted by Google/MIT and Cloudera both point to the same problem: Agents can’t reliably act on data they can’t access, understand, or retrieve in real time.

cio.com

Metabase SQLi exploit grants attackers total access

The critical zero-day can provide direct SQL access to Metabase’s underlying database, potentially exposing credentials, API keys, and other sensitive data.

csoonline.com

The future of AI security research isn’t autonomous, it’s human-amplified

HTTP Terminator generated, evaluated, and weaponized thousands of HTTP desync hypotheses, and surfaced a new technique that neither the AI or its human operator would have found alone.

csoonline.com

Cloudflare wants to provide the operating system for the AI-first enterprise

Cloudflare OS brings AI workspaces, zero-trust access, governed connectors, model routing, and enterprise data together in the browser, while new tools help IT teams track AI identities, token spendin...

cio.com

AI coding agents are blowing through budgets — Replit, Kilo Code, and Symbotic explain how they're managing it

Replit, Kilo Code, and Symbotic engineering leaders reveal how they track AI coding costs and stop runaway token spend before it wrecks the budget.

venturebeat.com

Ruby on Rails critical bug puts every image upload under scrutiny

The vulnerability abuses Active Storage, allowing unauthenticated attackers to read sensitive files and potentially take over applications.

csoonline.com

AI agents get better at IT ops, but only with humans in the loop

A study of nearly 150,000 real-world agent actions found AI increasingly handling low-risk IT tasks while human analysts iteratively shape and improve their actions.

computerworld.com

Microsoft doubles down on multi-model AI as it builds a Copilot super app

The company plans to combine chat, agents and business workflows into a single AI workspace while separating memory, context and orchestration from any one foundation model.

computerworld.com

Instacart's CTO says AI made the company stop worrying about tech debt

At VB Transform 2026, Instacart CTO Anirban Kundu said AI-generated code gets rebuilt so often that tech debt has stopped being a concern for his team.

venturebeat.com

5 startups tackling the AI agent trust gap | VentureBeat

One startup said it cut cyberattack containment time from seven hours to twelve minutes. See four other approaches to running AI agents safely at scale.

venturebeat.com

CISA unveils a six-step blueprint for isolating critical infrastructure during cyberattacks

A new guide from CISA and allied cybersecurity agencies says organizations should build dedicated isolation points into operational technology to contain attacks and continue delivering essential serv...

csoonline.com

Target SVP says its real AI moat isn't the models — it's everything built around them

Target SVP Siobhán Mc Feeney says AI models aren't the company's real edge — the governance layer it built so every agent has to earn its autonomy is.

venturebeat.com

A 13-year-old flaw is exposing tens of thousands of data center management systems

Lava researchers found more than 36,000 internet-exposed baseboard management controllers vulnerable to a 13-year-old IPMI flaw, giving attackers a path beneath the operating system.

csoonline.com

Hackers are compromising hotel Wi-Fi gateways to hijack Microsoft 365 accounts

By compromising captive Wi-Fi gateways instead of user devices, attackers can silently redirect authentication traffic and steal Microsoft 365 credentials.

csoonline.com